利用資料探勘技術探討台灣壽險市場發展之研究

Study on Development of the Taiwan Life Insurance Market Using Data Mining Techniques

劉介傳(Chieh-Chuan Liu)

指導教授 ： 詹勳生

嶺東科技大學/資訊學院/資訊科技應用研究所/碩士(2013年)

https://doi.org/10.6823/LTU.2013.00020

摘要

台灣保險業者在時間與經驗的累積，加上資訊科技的進步，紛紛將客戶資料導入客戶關係管理(CRM)系統，藉由客戶的基本資料及採購紀錄地建立，進行售後服務、商品規劃、保費收繳及出險理賠等工作，藉以維繫保險公司與客戶之間的關係。根據統計，70%保險新契約是來自原客戶「再購」，未來這些「原客戶」是否能再次購買保險，一直以來都是壽險公司及業務人員是否能成功經營的關鍵所在。然台灣目前人口相較10年前，已呈現負成長趨勢，少子化問題使得未來人口結構必趨於老化並衍生勞動力減少、扶養負擔加重、家庭結構及照顧、長期照護與社會保險等諸多問題，老人扶養勢將成為年輕世代的重大負擔，因此保險勢必成為更熱門的焦點。此外，在知識經濟、變動快速、競爭激烈的產業環境下，如何提昇競爭優勢以求有良好的組織績效表現，是現階段一項重要的課題。
本研究以抽樣問卷方式建立模擬資料庫，並運用SPSS工具導出資料庫之變異數與回歸分析，再利用資料探勘中關聯規則與決策樹技術，進行比對與驗證，期能挖掘出有價值的資訊，以利保險業者在決策與行銷面上能更精確並具競爭力。經研究實驗證明，透過問卷所建立之模擬資料庫除達信賴水準外，關聯規則方法可獲得數量變項與保險商品之偏好關係；決策樹方法則可獲得數量變項對保險商品購買意願，兩種方法經證明得到一個不錯的預測結果，可提供業界作為經營、行銷策略制定與商品設計的參考。

關鍵字

壽險 ； 客戶關係管理 ； 資料探勘 ； 關聯規則 ； 決策樹

並列摘要

With the accumulation of insurers’ experiences and the improvement of information technology, Taiwan insurers have applied customer relationship management (CRM) system to maintain the relationship between insurance companies and their customers by means of establishing personal information and purchase history of customers, after-sales service, product planning, premium collection, accident claims and so on. According to the statistics, 70 percent of new contracts were from "re-purchase" customers. The decision whether or not those old customers would re-purchase new insurance contracts has been critical when running insurance companies or for successful insurance sales. However, lower fertility in Taiwan is the major cause of an aging population, which leads to the problems of negative population growth, decrease of labor, increase of dependency load, change of family structure and care, and demand of long-term nursing and social insurance. In other words, senior caring will become great pressure to the young. Under this circumstance, the insurance industry should improve its service to gain better profit.
In the present study, questionnaire sampling was used to build a simulation database. With the use of variance and regression analysis of SPSS Software, data were analyzed, compared, and verified by means of the association rules and decision tree technology. The results indicated that the simulation database was relatively reliable. By using the approach of association rule, the preference of product and purchase number of insurance was correlated. By using the approach of decision tree, it was found that the number of insurance purchase may influence the willingness of purchase to some extent. These two methods provided a useful prediction and are suggested to be used as a reference of management, marketing strategies, and product design in the insurance industry

並列關鍵字

Life Insurance ； Customer Relationship Management (CRM) ； Data Mining ； Association Rule ； Decision Tree

參考文獻

一、中文部分：

1.中華民國人壽保險商業同業公會，「人身保險業務員資格測驗統一教材」，2010。

2.尹相志，「SQL Server 2008 Data Mining 資料採礦」，台北市，悅知文化，2009。

3.方慧臻、李宏安，「消費者對購買人壽保險行銷通路偏好之研究：以台中巿民為例」，壽險季刋，第134期，2004。

4.王保進，「視窗版SPSS與行為科學研究」，第二版，台北市，心理出版，2002。